The initial phases of these studies involved the in vivo central administration of neuropeptides on gastric acid secretion. The high affinity mu selective opioid heptapeptide dermorphin produced marked suppression of the secretion of gastric acid in conscious rats. This effect was blocked by the prior administration of indomethacin, which suggests that the synthesis of prostaglandins is required for the inhibitory action of neuropeptides on gastric secretion. These data are in accord with clinical studies which showed that individuals being treated with morphinomimetric analgesic drugs exhibit a lower incidence of ulcers in the mucosal lining of the stomach. These results led to the second phase: receptor analyses of the interaction of dermorphin with brain membrane binding sites. The binding data provided information on the formulation of a new model for the receptor interaction with selective opioid peptides: (1) deletion of Tyr5 or replacement by a hydrophilic residue, disrupted recognition at both mu and delta receptor sites; (2) blockage of the functional groups on Tyrl similarly reduced binding; (3) addition of hydrophobic protective groups on Ser 7 increased the affinity to mu and delta sites, but decreased mu selectivity; and (4) the requirement for a D configuration about the a carbon of residue 2 and amidation of the C-terminal residue is essential. This model proposes that the mu receptor accommodates two binding pockets: a T(mu) site for the stacked Tyr1/Tyr5 residues and a P(mu) site for the Phe 3 residue, which extends outward from the peptide backbone. Reversal of the Phe3-Gly4 dipeptide region in dermorphin to the Gly3 -Phe4 sequence in enkephalin drastically diminished mu selectivity. Thus, the N-terminal sequence, H-Tyr-D-Ala-Phe-Gly, particulates in a beta-turn through internal hydrogen bonds and constitutes the message domain of the peptide.